In the 2,500 years since Hippocrates, much has changed in the treatment of illness.1 The development of public health and hygiene, evolution of diagnostic techniques and imaging, and advances in the clinical armamentarium have all enabled clinicians and their patients to achieve remarkable improvements in life expectancy as well as quality of life. In all of these technological and scientific advances, however, one key aspect has not changed much since the age of Pericles and its most revered physician: the role of empiricism in determining the best medication for an individual patient.
One can imagine with little difficulty our predecessors saying the same as we do today: “Try this and let me know how you do.” Whether the patient was to inform the physician by smoke signal or Twitter, much of one’s treatment was, and remains, essentially predicated on an empiric model. We have come a long way in our understanding of therapeutic interventions. But, with regards to the medical management of pain, we have made only modest gains from the time of Hippocrates in predicting a priori whether a positive or negative response will occur. In pain medicine, this concept has been described as “N of one trial.”
This edition of Practical Pain Management has a fascinating, insightful, and not unexpectedly an “ahead of the curve” article by Forest Tennant, MD, DrPH. Dr. Tennant has written on the results of genetic testing performed on his patients suffering with intractable pain. Utilizing a relatively new approach to genomic testing, Dr. Tennant studied the incidence of metabolism variability in three key cytochrome P450 (CYP450) isoenzymes in patients on high doses of opioids (greater than 150 mg equivalents of morphine). The results were impressive: a majority of patients requiring high doses of opioids had some variation in CYP450 metabolism identified in his study. As a result of these findings, Dr. Tennant has proposed guidelines for testing certain pain patients (ie, those on doses cited in the article). Dr. Tennant shows us the way as he dissects out and identifies yet again an area sitting before our eyes, unseen and underappreciated.
Phamacogenomic concepts have been evolving since the ’90s and into the 2000s, with additional testing becoming available over the recent decade.2 It is one area of genomics that in all likelihood will have dramatic impact in the future on healthcare—covering a wide range of issues including diagnosis, treatment, and prognosis.3,4 For instance, genetic markers of disease are routinely being used in a number of conditions as drivers of—and predictors for—which specific therapies may work, as well as anticipating response to those treatments. Currently, testing drug response and metabolism are being used in different clinical contexts.5 Examples include CYP-2D6 testing in patients suffering with breast cancer to determine whether tamoxifen can be used. Tamoxifen is a pro-drug and requires biotransformation to the active metabolites afimoxifene and endoxifen to be effective. Absence of or reduced activity in 2D6 enzymes either precludes or may require different dosing. Another example is clopidogrel (Plavix). Clopidogrel is also a prodrug requiring appropriate metabolism through the CYP-2C19 enzyme system. In March 2010, the FDA put a black box warning on clopidogrel to make patients and healthcare providers aware that CYP-2C19 poor metabolizers, representing up to 14% of patients, are at high risk of treatment failure and that testing is available.6 Researchers have found that patients with variants in CYP-2C19 had low levels of the active metabolite of clopidogrel, which resulted in less inhibition of platelets, as well as a 3.58 times greater risk for major adverse cardiovascular events such as death, heart attack, and stroke. More than 70 drugs—or greater than 10% of FDA-approved medications—now have some reference to pharmacogenomics on their labels.7
For many clinicians, interpatient and intrapatient variability of response and tolerability to opioids have proven difficult concepts. Those who regularly see and treat pain patients know that these varied responses do exist, but have lacked clear and scientific basis to document possible causes or contributors. The breakthrough work done by Pasternak et al have demonstrated variability in laboratory animals’ responsiveness to opioids and may help explain some of the variability to response seen. One potential area of interest to the pain practitioner is the study of the OPRM-1 (µ opioid receptor) gene, which is putatively responsible for opioid responsiveness and is a possible marker for certain populations at risk for alcohol and opioid abuse. However, this research has not yet been validated in clinical practice. Another area of interest and controversy is the role of testing catecholamines and serotonin through the catechol-O-methyltransferase gene.8,9
Now, however, aspects of pharmacogenomics—particularly in the use of opioids and their role in treating pain—may be ready for more regular uses in our clinics.8 Currently, testing is commercially available to provide insight into several key areas affecting our plans of care. These include the role of CYP450 testing for certain genetic traits that determine drug metabolism, such as biotransformation of prodrugs and the formation of active metabolites. This information may lead to a clearer understanding of benefits as well as risks of drug–drug interactions. Dr. Tennant’s work suggests that in his population, a high prevalence of varied metabolism exists as a result of genetic variability in patients tested.
Based on the work presented in Dr. Tennant’s article, it is clear that variability in drug metabolism may be higher than expected. What is unclear is how this will apply to daily practice. Personal experience corroborates Dr. Tennant’s findings concerning a higher than predicted rate of variant CYP450 expression, particularly in CYP-2D6 expression. But, does the work of pain clinics treating high-dose patients translate to the greater population at large or pain patients in particular? Because we can test for something, does it mean we have to?